Mechanisms
The development of microstructure is considered to proceed by a basic three- step process (Isenburg, 1974; Ohama, 1987). Immediately after mixing, the system consists of well dispersed un-hydrated cement particles, aggregate, and polymer particles. After the dormant period, cement hydration starts to accelerate and, as water is consumed, the inter-particle separation distances decrease. As a result of this, the latex becomes unstable causing the polymer particles to flocculate and deposit on the C-S-H and aggregate surfaces. With further removal of water, the polymer particles coalesce to give a continuous three-dimensional film or membrane. This membrane binds the cement hydrates and aggregate and also improves the bond between the polymer modified cement and existing concrete or steel substrates. It has also been suggested that any ionised carboxyl groups will bond to cement grains via calcium ions forming `chemical anchors’, see Fig. 10.1(c), (Chandra, 1987; Ohama, 1998).
To achieve this type of microstructure, a conflicting set of curing conditions is required. Cement hydration needs water to give the necessary hydration products and the development of compressive strength. In contrast, the polymer phase requires dehydration so that the particles coalesce to form a continuous film, which is essential for improvements in flexural strength and toughness. Thus a two-stage curing regime is often recommended consisting typically of 24 or 48 hours `wet’ followed by 28 days `dry’.
The evidence for this mechanism stems from a publication based on work carried out many years ago (Isenburg, 1974) where a distinct polymer film bridging a crack was observed by scanning electron microscopy (SEM). In practice it is very difficult to elucidate the true nature of the polymer films, although modern instruments have facilitated observations on the micrometer scale for a wide range of polymer systems (Su, 1996; Ollitrault-Fichet, 1998; Jenni, 2002). Sample preparation is important and observations may be made of fracture surfaces or cross-sections, with or without etching of the cement phases (Zeng, 1996a). In some areas the film is present on a very fine scale and exists as a discontinuous mesh or in a fibrous form which is intertwined with cement hydration products. In other areas the film is more distinct bridging pores and micro-cracks. The films formed when the polymer is added as a redispersible powder have been compared to those formed when the polymer is added as latex (Afridi, 2003). It was found that the former have an inferior quality compared to the latter, which was attributed to a less uniform polymer distribution in the mix and poor coalescence of the polymer particles during curing.
Kinetics
It is considered that the addition of polymer latexes will generally retard the hydration of cements. Isothermal conduction calorimetry has been used (Zeng, 1996b) to study the hydration kinetics of SBR latex modified cement. The effect of increasing amounts of SBR on heat evolution is shown in Fig. 10.5. It is clear in this case that the SBR acted as a retarder of cement hydration in that it increased the induction period and time to maximum heat output whilst reducing the total heat outputs. Low concentrations of SBR (<10%) had a significant effect during the acceleration period, reducing the rate of alite reaction and degree of hydration by about a half. Higher concentrations (>10%) suppressed the nucleation of cement hydration products and thus increased the induction period, with only small further reductions in the rate of alite reaction and degree of hydration.
Calculated activation energies for the hydration of unmodified and modified cements containing up to 10% polymer were very similar and suggest that the rate controlling step, diffusion of ions through a reacted layer, was unaltered in the presence of polymer particles. That the actual rate is reduced suggests that polymer particles are occupying some of the sites (see Fig. 10.1(c)) available for calcium dissolution and subsequent hydration. This has been confirmed by SEM observations (Su, 1996). In the case of the modified cements containing 20% polymer, the activation energy was much lower than that for the unmodified cement. This suggests that a different mechanism is now controlling the rate of hydration, which may involve restricted movement of water to hydration sites. That the actual rate is low may be because the configuration of polymer particles is such that they provide a more tortuous path for water diffusion, e.g. through a sheath formed by the polymer around the cement particles.
Whilst the constituents of the aqueous component of the latex had little influence on kinetics, an increased concentration of carboxyl groups in the polymer significantly increased retardation. It is thought that this was a result of increased binding between the carboxyl groups and hydration sites on the surface of the cement grains.
It is evident from this work that polymer-cement interactions are important since they have the ability to influence hydration kinetics and thus in turn influence porosity and polymer film development, both of which have a long- term effect on durability.
Porosity
The influence of polymer modification on the total porosity and pore size distribution of cement pastes has been studied by mercury intrusion porosimetry (Zeng, 1996a). Whilst there are clear reservations in using this technique (partly for the reasons indicated in Chapter 2 and also because of the compressibility of the polymers) results were checked by comparing values of total porosity with those obtained by solvent exchange, water absorption and helium pycnometry. This showed that the different techniques resulted in different absolute total porosity values, but trends with, e.g. curing time, curing condition and polymer type and content were essentially the same (Zeng, 1996a).
Wet curing of a PMC tends to result in an increase in total porosity and give a coarser pore structure compared with similar unmodified cement, at constant w/c. This may be attributed to the retarding effect of the polymer although the effect is reduced over time so that, with longer curing, e.g. 90 days, little difference is found. Real reductions in porosity come from the ability to use lower w/c, whilst at the same time maintaining the workability.
in improved mechanical properties, such a regime is contrary to that which would be expected for a low porosity surface layer (Cather, 1994) e.g. desirable, for good resistance to chloride ingress. In other than water-saturated curing, it is thought that the water loss will be most rapid at the outer surface and lowest at some point remote from the surface. However, it has also been suggested that the rate of evaporation from polymer modified cements is lower compared with that from unmodified cements (as determined by mass loss for the total sample) and that this decreases with an increase in latex content (Ohama, 1982). Pore size distribution curves for wet-dry cured unmodified and 10% and 20% SBR modified cement paste at w/c 0.35 are shown in Fig. 10.6 (Salbin, 1996; Short, 1997). Samples were taken from the first 2 mm of surface and at depths of 15 and 30 mm from a prism with one exposed surface, which had been subjected to a curing regime of one day wet and 27 days dry. In the case of the unmodified cement, there is no distinct difference between the pore size distributions at 15 and 30mm depths, total porosity and initial pore entry diameter being about 0.11 cc/g and 0.20 um respectively. These values are consistent with those obtained for specimens cured for 28 days wet. However, the surface layer has a coarser pore structure with a total porosity of 0.14 cc/g and initial pore entry diameter of 1.9um This difference is presumably a result of insufficient hydration through evaporation of water from the surface layer. Thus a curing affected zone (CAZ) exists although it does not extend to 15mm from the surface. In cement modified with 10% SBR, the surface layers are much coarser than the middle and bottom layers and furthermore the layers are much coarser than in the case of those taken from unmodified cement. It would appear that, for this SBR, the wet-dry curing regime had an adverse effect on pore structure. Increasing the polymer content to 20% SBR results in slightly less coarse pore structures compared with specimens containing 10% polymer. They are, however, still coarser than those of the unmodified cement specimens. It may be that greater polymer contents reduce evaporation of water to some extent, but with this particular polymer the effect is not very significant.
In the case of Acrylic and Ethylene Vinyl Acetate modified cements, changes in pore structure as a result of the wet-dry curing regime were not as severe as in the case of SBR, although a CAZ still exists. The reasons for this difference are not clear and this emphasises the need to appreciate that different commercial systems may behave in different ways. Investigations should concentrate on moisture movement at the surface of specimens and not in the bulk of the specimens as a whole.
